One vibe to rule them all - Unify your AI coding assistants with shared standards
Your team uses different AI tools on the same codebase:
- π Sarah uses Cursor (needs
.cursor/rules/*.mdc) - π€ Tom uses Claude (needs
CLAUDE.md) - π Mike uses Copilot (needs
.github/instructions/*.md)
Without vibe: Sarah improves the React hooks guidelines. Tom and Mike never see these improvements. Three developers, three different versions of "truth" = chaos.
With vibe: Sarah updates vibe/react-patterns.md. Everyone gets the improvement automatically. One source of truth = harmony.
# 1. See what your project needs (10 seconds)
npx @jinjos/vibe-cli analyze
# 2. Set up vibe
npx @jinjos/vibe-cli init
# Done! All your AI tools now share the same rulesπ REPOSITORY ANALYSIS REPORT
==================================================
π Tech Stack Detected:
Frontend: React, Next.js
Database: MongoDB
Testing: Playwright
π€ AI Configuration Maturity: 21/100 β οΈ
β MISSING RULES FOR YOUR STACK:
β’ MongoDB: connection patterns, indexing, schemas
β’ React: hooks guidelines, performance patterns
β’ Security: input validation, auth patterns
β’ Error handling: logging, recovery strategies
π― NEXT STEPS:
1. Create MongoDB connection rules
2. Add React component patterns
3. Run: vibe init
Why this matters: Instead of guessing what rules to write, vibe analyze tells you EXACTLY what your project needs based on your actual tech stack.
Add to your Claude/Cursor MCP config:
{
"mcpServers": {
"vibe": {
"command": "npx",
"args": ["-y", "@jinjos/vibe-mcp"],
"cwd": "/path/to/your/project"
}
}
}npm install -g @jinjos/vibe-clivibe analyze # Get AI configuration health check (0-100 score)Shows exactly what rules your tech stack needs.
vibe init # Creates vibe/ directory and configures all platforms
vibe init --full # Also migrates existing rulesvibe sync # Pull rules from platform directories to vibe/
vibe sync --full # Sync and remove originalsvibe status # Quick health check
vibe status --verbose # Detailed performance metricsBefore vibe:
project/
βββ .cursor/rules/react.mdc # Sarah's version
βββ CLAUDE.md # Tom's version
βββ .github/instructions/ # Mike's version
βββ react.md
After vibe:
project/
βββ vibe/ # One source of truth
β βββ react-patterns.md # Shared by all
β βββ mongodb-rules.md
β βββ security-standards.md
βββ .cursor/rules/vibe.mdc # β Points to vibe/
βββ CLAUDE.md # β Points to vibe/
βββ .github/instructions/ # β Points to vibe/
βββ vibe.instructions.md
| Platform | Context Window | Config Location |
|---|---|---|
| π― Cursor | 200K tokens | .cursor/rules/vibe.mdc |
| π€ Claude | 200K tokens | CLAUDE.md |
| π Copilot | 64K tokens | .github/instructions/vibe.instructions.md |
| π Gemini | 1M tokens | GEMINI.md |
vibe analyze # See what you need
vibe init # Set it up
# Create the rules identified by analyze
vibe status # Verify everything works# After teammate adds rules to .cursor/rules/
vibe sync # Pull their rules into vibe/
git commit # Share with teamvibe status --verbose
# Output:
β
CURSOR: 1,580 tokens (1% of limit) - Good
β οΈ COPILOT: 1,580 tokens (2% of limit) - Monitor- Know Before You Code -
vibe analyzeshows gaps before writing rules - Single Source of Truth - No more copy-pasting between platforms
- AI Health Score - Track your configuration maturity (0-100)
- Tech-Specific - Get recommendations for YOUR stack, not generic advice
- Zero Config - Works out of the box
MIT
π₯ Give your AI tools the perfect vibe π₯